Modern computing devices have become increasingly reliant on network-based content and services to provide an enhanced user experience. As a result, host computing devices that provide network-based content and services need to be available to answer requests received over a network. In order to provide such availability, host computing devices are often configured to operate in an always-on and always-connected manner. However, due to raising energy prices and an increasing number of “green initiatives,” many network-based content and service providers are utilizing a variety of energy-saving, power-managed states that are typically built into host computing devices such as, for example, low-power sleep or hibernate states.
To increase the availability of network-based content and services, providers may configure host computing devices to remotely wake up upon receiving one or more specially-identified packets, subsequently referred to as packets or wake packets, over the network from a requesting device. Typically, the requesting device expects a reliable connection to be formed with the host computing device shortly after sending a packet to initiate a connection. Unfortunately, it often takes some amount of time for one or more components of the host computing device to reach an operational state after receiving an initial packet that triggers a wake-up. This is especially true in complex host computing devices having a number of interdependent components. As a result, the initial packet is often lost and the requesting computing device is typically required to resend the packet at increasing time intervals, which in turn, increases the overall time needed to establish a connection.